Sprint - Jan 6 - Jan 17

[marandaneto]

Global Sprint Planning

3 things that might take us down

Urgent incident follow-ups

https://github.com/orgs/PostHog/projects/103/views/2

Team sprint planning

For your team sprint planning copy this template into a comment below for each team.

# Team ___

**Support hero:** ___

## Retro

<!-- Grab the high and low priority items from last time and add whether that item was completed or not -->

- 

## Hang over items from previous sprint

<!-- For each item, decide to re-prioritise (and add below) or deprioritise -->

- Item 1. prioritised/deprioritise

## OKR

1. OKR, status (red/yellow/green) and action points if yellow/red


### High priority

-

### Low priority / side quests

-

[timgl]

For your team sprint planning copy this template into a comment below for each team.

Team Exporters of Batches

Support hero: Ross first week, Tomás 2nd

Retro

• Split file size shipped
• _inserted_at for TFH shipped
• Fixed the bigquery quota problem -> compression factor is 9x compared to before + we use stage tables. We can do 10x, maybe can't do 100x.
• Better alerting
• Cancelling batch export queries

Hang over items from previous sprint

• Item 1. prioritised/deprioritise

OKR

1. OKR, status (red/yellow/green) and action points if yellow/red

High priority

• Batch exports: Add key-pair for snowflake exports #26692
• Property filters for Batch Exports #26415 (use hogql in the backend to do this)

Low priority / side quests

• https://github.com/orgs/PostHog/projects/114/views/7?pane=issue&itemId=92837228
• encrypt

[Twixes]

Team AIPM

Retro

• ✅ Make summarizer self-aware when we couldn't fully answer the question, and why @Twixes
• 🟡 All other classic insights types (order of importance: retention, paths, stickiness, lifecycle) @Twixes
• ✅ Max AI rolled out as feature preview @PostHog/team-ai-product-manager
• ✅ Human-in-the-loop for clarifying taxonomy @skoob13

Hang over items from previous sprint

• Stickiness/lifecycle/retention insights

High priority

• Core memory, with UI to manage memories @skoob13
• Address user feedback:
  • Keep showing RAG actions that were taken after insight creation is done @Twixes
  • "Stop response" button @skoob13
  • Ability to type while assistant is generating @skoob13
• Enable LLM tracing in production for visibility (choice between Langfuse vs. Langsmith vs. Helicone) @skoob13
• Archival storage @skoob13
• Stickiness/lifecycle/retention insights @Twixes
• AI in "Add insight" on dashboards @Twixes

Low priority / side quests

• Fix empty message just before schema generation kicks off @skoob13

[daibhin]

Team Error Tracking

Retro

🟢 Q1 goals decided @daibhin
🟢 Launched early access beta (33 teams sending exception events!) @daibhin
🟢 Materialized columns for $exception_fingerprint and $exception_issue_id @neilkakkar
🟡 Postgres backed issue query to improve performance (#27157) @neilkakkar
🛑 Blocked on threshold based alerting by #27072

OKR

🟢 Migrate PostHog off Sentry

• Figure out why it breaks billing
• Get to the bottom of data discrepancy issues
• Implement automatic team based issue assignment

🟢 Ship the damn thing

• Cymbal metric dashboards & alerts
• Exclude $exception events from analytics billing
• Fix query performance to make UI snappier
• Implement issue threshold based alerting (time and volume)
• Add build time symbol set upload support

High priority

Migrate PostHog off Sentry

• Fix billing issue with upgrade @neilkakkar
• Assign issues to teams @daibhin

Ship the damn thing

• Ship new Postgres backed error tracking issue list @neilkakkar
• Sync up with Ben to finish off the last bits for alerting @daibhin
• Metrics dashboard for cymbal @daibhin

[marandaneto]

Team Surveys

• Support hero @marandaneto
• No PTOs

Retro

• 🟢 Customer support (some issues need a @dmarticus pair of eyes still)
• 🟢 Onboarding Lucas @dmarticus
• 🟢 Lucas ships 1st PRs @lucasheriques

OKR [TBD]

High priority

• Q1 planning @marandaneto and team
• allow admins to configure font in the survey editor @lucasheriques
• Define a plan for Mobile surveys - manual presentation vs Builtin presentation - PoC @marandaneto @ioannisj
  • Do we implement all features or just a subset of them? All display conditions, etc.
• Lay the foundation for Mobile surveys on iOS and React Native @marandaneto @ioannisj
  • HTTP requests - parsing responses, etc
  • Migrate the Decide API to the new endpoint (only call flags if needed)

Low priority / Side quest

• [ bug ] Wrong graph series after changing question order @lucasheriques
• [ improvement ] Use a bar chart instead of a trendline for the survey NPS insight @lucasheriques
• Finish up the network instrumentation refactoring for iOS session replay @ioannisj
• Stop and resume all integrations on opt-out/opt-in @ioannisj
• Survey targetting platform, eg Mobile only, Web only, etc @marandaneto @lucasheriques
• 🐛 bug Surveys is being showed multiple times @lucasheriques

[danielbachhuber]

Team Experiments

Retro

• 🟢 Finish hogql migration cleanup. @andehen
  • ~1000 lines of code deleted
• 🟢 Make it easy to generate experiment test data @andehen
  • A django management command generate_experiment_data now exists so one can easily generate test data PR
  • As part of this, ran lots of tests to get experience with the product and reflected on pain points
• 🟢 Continue working on the stats methods. @danielbachhuber
  • New stats methods are live in production and applied to experiments created after January 6th.
  • Published a few new docs.
• 🟢 Get started on the reusable metrics UI, have a minimal version rolled out for our priority customer. @jurajmajerik
  • Live for 8 customers with experiments-multiple-metrics feature flag enabled. Going live for all customers on Thursday.
  • Nice words from Sam @ Bingo Card Creator: "The new shared metrics and multiple metrics is great, this will save me a lot of time remaking funnels each time."

Sprint goals

Support hero: @andehen

@andehen

• Fix exposure bug that causes wrong results (not all users in the experiment are included)
• Improve query validation error response and UI
• Support

@danielbachhuber

• Mega documentation polish pass
• Try to get Fix "Filter internal and test users" when personsOnEventsMode = PersonsOnEventsMode.PERSON_ID_NO_OVERRIDE_PROPERTIES_ON_EVENTS #27132 to resolution

@jurajmajerik

• Catch up after holiday
• Refactoring, cleanup, and restoring tests

[robbie-c]

Team Web

Support hero: @robbie-c + @rafaeelaudibert

Retro

Main quest

Robbie
🟠 Finish cookieless prototype

Rafael
🟢 Core web vitals product research

Side quests & support:

Robbie
🟢 Custom channel types can use full URL and pathname #27226
🟢 Make posthog-rs send $lib_name correctly, and add some GitHub actions PostHog/posthog-rs#13
🟢 Slightly better support for mobile apps in web analytics #27242
🟢 Add custom bounce rate duration control #27121
🟢 Add pageview ID tracking to posthog-js PostHog/posthog-js#1634

Rafa
🟡 Human-friendly date conversion
🟢 Support SDK docs updates

OKR

• Finish cookieless mode
• Build marketing analytics
  • Add revenue data to web analytics
  • Get ad spend data into data warehouse and use it for per-channel analytics
• Build per-page reports
  • Core web vitals

Sprint goals

High priority

Rafael

• Core web vitals in beta
• Helping out with cookieless

Robbie

• Finish cookieless (no more prototype)

[benjackwhite]

Team Cambridge Damaged Plumbing (CDP)

• Support: @benjackwhite

Retro

• 🟢 Keep things stable and running!
  • Upgraded MSK successfully without any data issues 🥳
• 🟡 Finish the internal event source destinations @benjackwhite
  • Working for ActivityLog. Still work to do for Errors alerts
• 🟡 Rollout RemoteConfig when it feels very safe (or just get more metrics to rollout post-break) @benjackwhite
  • Started but found some edge case errors so rolled back
• 🟡 Tiktok / snapchat client side destinations - running on an actual website of some kind! @MarconLP (could chat with marketing team to see if there is a good thing we could test this with)
  • No tiktok yet waiting on mappings
• 🟡 Batch imports @oliverb123 - getting at least one of the target customers imported via this.
  • Getting there, holiday slow down
• 🟢 Be ready to onboard 2 new people!

OKRs

• 😍 People choose us for their CDP first
• 🖕 Plugins be gone
• 🚀 Warpstream for events
• 🎈 Write-optimized session replay ingestion
• 💪 We make changes and improvements to the Node services with confidence

High priority

• Get mappings work released @MarconLP @benjackwhite
• Finish the error issue alerts work @benjackwhite
• Rollout RemoteConfig @benjackwhite
• Migrate intercom plugins with confidence @MarconLP @meikelmosby
• Roadmap and scaffolding in place for next version of Blobby @pl (w support @benjackwhite)
• Migrate all action webhooks to Hog functions @meikelmosby
• Batch imports for the customers it is focused on @oliverb123

Stretch goal

• Capture mirror for Warpstream

[zlwaterfield]

Team Growth

Retro

Retro items

Pretty light with xmas break

• @zlwaterfield
  • lots of misc RBAC things for UI and edge cases
  • new Google SSO authentication logic
• @patricio-posthog
  • shipped new forecasting (p95 dropped 50% 🙌)
  • turned forecasting emails back on (first invoice and large change)
  • working with Joe on ElevenLabs credit programs
  • working through mobile pricing changes (not done yet)
• @surbhi-posthog
  • teams add-on usage events
  • worked through experiment for onboarding flow for non-techincal users
• @joshsny - joined the team 🙌

Q1 2025 Objectives

✅=finished 🟡=in progress 🔴=won't finish ⚪=not started

• 🟡 Usage dashboard (Pato)
• Pricing changes(Pato)
  • 🟡 Finish Mobile Replay pricing
  • ⚪ Query API pricing
  • ⚪ Clean up billing for identified / unidentified events
• 🟡 RBAC (Zach)
• ⚪ Enterprise add-on (Zach)
• ⚪ Improve SAML (Zach)
• ⚪ Use intent data to onboard users into multiple products (Surbhi and Josh)
• ⚪ Surface intent data later in the users journey (Surbhi and Josh)
• ⚪ Move core RevOps flows off of Zapier into the billing system (tbd)

Extras (nice to haves if we have time or feel inspired)

1. Postgres (pgbouncer + read replica) - infrastructure handling this
2. Self serve account deletion
3. Case insensitive emails
4. Move all transactional emails to Customer.io
5. 3D Secure improvements - email flows
6. Move revenue dashboards from Metabase into PostHog
7. Get credit cart input experiment over the line
8. Billing endpoint performance improvements (split up)
9. Proper callbacks for time-sensitive checks

This sprint

Time off: @patricio-posthog off for 8 days

• @zlwaterfield
  • RBAC
    • client side checks
    • internal/external documentation
    • first attempt at a migration
    • talk to each product team about implementation details
    • organization member permission changes
  • Onboarding @joshsny
  • Customer.io transactional emails in posthog (if time) and 2FA emails (if time)
  • Billing pg-bouncer (w/ infra)
• @patricio-posthog
  • Keep on the analysis for January forecasting
  • Continue with the Mobile replay new pricing
• @surbhi-posthog
  • Support hero this week
  • Working on shipping first onboarding experiment this week
  • Closing out team logging events related to white-labeling and SSO enforcement
• @joshsny
  • First week - onboarding and next week in SF
  • Improving some styling through onboarding process
  • Using fresh eyes to ideate on improvements for onboarding/activation and cross sell opportunities

[mariusandra]

Team Devilishly Exciting

Retro

• Adam: how to go about implementing a design system. Implementing in the app, new navbar. Gone through every element in the top bar, replacing with a new accessible component (keyboard shortcuts, etc)
• Marius: pnpm workspaces and xmlsec upgrade
• Cory: "all products" view inside the app

OKR

• Build a framework where we can have 50+ products? @mariusandra
  • Defining and splitting the common layer from the rest of the apps
  • Defining and implementing boundaries between apps
  • UX changes to support the new structure
  • Every product is able to export operations (backend) and widgets (components) => AI, notebooks etc.
  • Testing independently - if you change things on one product, we only run tests for that product
• Levelling up our product quality. @adamleithp
  • Design system that is self documenting
  • Mobile support
  • Finally figure out (explain) modals vs drawers
  • More standardised components
• Anti-goal for Q1: @corywatilo
  • Fixing CI when it randomly breaks (each team owns their tests)
  • Rewrite everything
  • Implement everything ourselves

High priority

• Adam: keep working on components in the shadows, design system work. First components: top nav, dropdown menus, tabs and other generic layout features.
• Marius: project folders - reorganize our codebase to support the next 50+ products
• Cory: keep iterating on the design concepts

Low priority / side quests

• Marius: Hog/QL support

[MichaelKutsch-ph]

Team Infra

Support: @frankh

Retro

• 🟡observability for internal-cluster - Kafka/RDS incidents blocked this @MichaelKutsch-ph
• 🟡 iterate on cloud spend topic - Kafka/RDS incidents blocked this @MichaelKutsch-ph
• 🟡 argoCD alerts: prioritized manual deployments for now as this is more important for day-to-day work for incident response @bciaraldi
• 🟢 we won't be doing big things as we're mostly off anyways and not breaking something because so many peeps are off as well- we upgarded MSK to Kafka 3.7. :D
• 🔴 dropping: more VPC CIDRs for scaling concerns @frankh - we have other pressing issues, so we postpone that one for e.g. move to WS
• Many incidents in the last weeks (MSK, RDS Indexes...) do not feel that good - we need to tackle things in this area especially considering growth plans. Upgrading MSK was one first step, we have to follow up here and also tackle the big mothership (holy postgres of death)

OKR

1. 🏃 Next level ArgoCD - Continuous Delivery
2. ☁️ 🕸️ Scalable networking
3. 💰 Mr. Bezos, bring down those costs!
4. 🎮 A stable pre-prod environment

High priority

• AWS PPA @MichaelKutsch-ph
• Kafka failover is safe @MichaelKutsch-ph @frankh
• RDS - incident follow ups (tbd after review meeting) - whole team
• runbook for argo operations for teams @bciaraldi + demo in all hands
• continue argoCD alerts @bciaraldi
• Get a argoCD workflow going for great DX for teams @danielxnj
• probably Warpstream @frankh
• continue observability for internal-cluster @MichaelKutsch-ph

Low priority / side quests

[pauldambra]

# Team Re(tention)play

R – Recordings
E – Erasing churn
P – Patching rrweb
L – Listing for you
A – asking user questions
Y – Yippee-ki-yay!

xmas sprint was quiet... support is pretty high now as a result

Support hero: @pauldambra

Retro

## what we shipped

Mobile

• 🔄 Capture network headers and body (assuming new instrumentation is in place by then) Network bodies for network logs posthog-ios#159 @ioannisj
  • deprioritised to get the other things out - carried forward
    ✅ move mobile replay from jpeg to webp
  • android is out, iOS soon to follow
  • aim is to see proportion of overall data coming from mobile snapshots falls
    ✅ manual start and stop of recordings on all platforms so users can implement whatever ingestion control they need

Replay

🚫 brandy butter
✅ brandy cream

OKRs 🟡

1. OKR, status (red/yellow/green) and action points if yellow/red

🟢 pending... PostHog/posthog.com#10197

High priority

• update to rrweb 18 including getting dialog element support into front end @pauldambra
• get support back under control after xmas @pauldambra
• user interviews around "how to get value out of recording" @veryayskiy
• default playlists - HogBot creates playlists for everyone like templates/insights @veryayskiy

[EDsCODE]

Team Data Storehouse

OKR Q1 2025

Objective

Objective: World class query experience

• Key Results:
  • SQL copilot
  • Draft mode
  • Data modeling and lineage controls

Objective: Data Visualizations

• Key Results:
  • Complete suite of charting options
  • External data for all insights

Objective: Extensible sources

• Key Results:
  • Repeatable and easy to build data import source pattern
  • Revised source configuration experience

Retro

• data modeling visualization in editor released internally (behind a flag)
• Draft queries + reorg sidebar (Epic: Editor & Modeling 3000 Epic: Editor & Modeling 3000 #26193) @EDsCODE
  • released as early access features
• Data import deployment and testing @Gilbert09
  • memory footprint reduced a ton (~50gb -> <10gb). There's still what seems to be a leak though with the vacuum and compaction job

Priorities

Tom off second week

• onboarding @EDsCODE
• sidebar folders @EDsCODE
• Pipeline v2 @Gilbert09

[fuziontech]

Team ClickHog

Objective

Goals for Q1:

• Query Scheduling Service
  • Paid API use case is covered
  • We should have all queries running through this
  • Basic load shedding
  • Differentiate API queries based on $
• Iceberg continuation
  • Keep on top of Altinity builds for CH
  • Query logs into Iceberg
  • Ingestion pipeline work (monitoring & automation)
• Cluster topology consolidation and cost reductions
  • We are on one cluster
  • Query performance is equivalent
  • Costs are down
  • Schema management
• Ongoing HogQL/Performance improvements
  • Query Scheduling
  • Productizing Performance Tooling
  • Better query attribution (in query comments)
    • By end of query we do not run queries with no annotations
    • Some sort of Type for queries that will run

Board

https://github.com/orgs/PostHog/projects/85/views/2

Retro

🔥 Incidents 🔥

Last Sprint's Goals

Not much - KTLO ideally

• Benchmark new cluster @fuziontech
• Shut down new cluster @fuziontech
• Get into a place where we can quickly restart new cluster @fuziontech
• chproxy into dev stack @orian
• Look into weekend event delete job @orian
• Look into iceberg for query log @orian
• Squash job @tkaemming

This Sprint's Goals

TL;DR

• Add operational observability to HouseWatch @fuziontech
  • Replication exceptions and postpone
• Tiered storage TTL strategy (looking into query partition distribution) - A bit of everyone
• Switch Django app to use HTTP clickhouse client (for CHProxy support) @orian
• Getting CHProxy in Dev @orian
• Squash workflow starting to run in prod @tkaemming
• Look into exposing metrics through Prometheus Format that expose over some /metrics endpoint @Daesgar
• Schema management with HousePlant @Daesgar

[aspicer]

Team Product Analytics

Q1 2025 objectives

1. Main Goal: Make it a no-brainer for customers to use our insights

   • Reduce usage of HogQL as a workaround/crutch for missing UI features
   • Research and improve insights
     • Funnels
       • Filters in exclusion steps
       • Optional Steps
       • Time to convert trends
     • User Paths
     • Retention
       • Revenue Retention - connect to data warehouse
     • Data management / Non-existing tools
   • Better path cleaning support
   • Taxonomic Filter Improvements
     • Better organized, easier to use
     • Information about properties (schema and statistics)
   • Make other non-event sources a first-class citizen of the insight ecosystem

2. Small Goal: Alerts v2

   • Alerts on funnels
   • Notifications through Slack

3. Small Goal: Shared dashboards with passwords

4. Research Goal: Understand activation

Sprint Plan

Support hero: @thmsobrmlr @anirudhpillai

• Finish open topics
  • Legacy API conversion to HogQL [Thomas]
  • Personless events actors modal fixes [Thomas]
  • Customizable colors in insights follow ups [Thomas]
  • Run migration of insights from filters to query [Thomas]
  • Dashboard creation / edit experience improvements [Peter]
  • Dashboard Filters and Variables in URL [Peter]
  • Deploy and monitor issues with hogql cohort calculation [Sandy]
• Insight Research and Planning

[dmarticus]

Team Fast Lightweight Accurate Good Stuff (Flags)

Retro

Short sprint, didn't get too much feature work done (outside of Haven finishing feature management CRUD ops)

• Finished basic feature management CRUD ops @havenbarnes
• helped out with Lucas onboarding for team surveys @dmarticus
• added more metrics for production testing for new flags @dmarticus
• prepared onboarding for @haacked !

Q1 2025 Objectives

(I'm going to turn these into sub-issues so I don't have to type so much)
✅=finished 🟡=in progress 🔴=won't finish ⚪=not started

🟡 Performance & Reliability

• 🟡 Complete migration of feature flag evaluation traffic from the legacy /decide endpoint to the new /flags endpoint (@dmarticus)
• ⚪ Establish comprehensive monitoring and observability for the feature flag platform (@dmarticus, @havenbarnes)
• ⚪ Enhance Grafana metrics to differentiate between infrastructure failures and logical errors (@dmarticus)
• ⚪ Collaborate with infrastructure team to create infrastructure cost analysis and dependency mapping with Michi (@dmarticus)
• ⚪ Improve observability of the feature flag evaluation process by tying flag evaluation results to the original request context (@havenbarnes)
• ⚪ Experiment with sampling request parameters for high-value customers to have better observability for support requests (@havenbarnes)
• ⚪ Implement actionable Sentry alerts in #alerts-feature-flags channel (@havenbarnes)
• ⚪ Integrate PostHog error tracking with new flags endpoint (@dmarticus)
• 🟡 Implement read-only flags (specifics to be defined, see the following RFC) (@havenbarnes)

🟡 Customer-Driven Innovation

• ⚪ Behavioral cohort targeting for flags (RFC here) (@dmarticus)
• 🟡 Feature management (@havenbarnes)
• ⚪ Automatically manage flag staleness (@haacked)
• ⚪ Customer research-driven features (TBD based on the findings of the market research and customer interviews)

⚪ Strategic Research

• Identify and analyze key user segments:

1. Current active users
2. Competitor users
3. Custom solution implementers
4. Potential users unaware of feature flagging

• Conduct customer interviews to identify gaps and growth opportunities
• Document findings and recommendations for product roadmap.
• More context here

🟡 SDK Spike

• ⚪ Evaluate feasibility and tradeoffs around adding feature flags to the Java SDK
• ⚪ Investigate improving the Rust SDK (add feature flag evaluation, implement the suggestions from the open issues etc.)
• 🟡 Research .NET SDK demand + feasibility @haacked
• ⚪ Fix open issues in the Python and Node SDKs: (Python, Node)

🟡 User Experience Improvements

Low-hanging, medium-impact, and largely context-free fruit that can be picked up by any team member during available capacity:

• ⚪ Consolidate boolean and multivariate flag interfaces
• ⚪ Show flag evaluation results more clearly in the flag editor UI
• ⚪ Improve the experience of syncing flags between environments
• 🟡 Anything else from the current feature flag backlog

This sprint

Themes: onboarding (Phil and Lucas), new flags production readiness, feature management chugging along

Time off: @dmarticus off for 1 day, @haacked off for 1 day.

• @dmarticus
  • New flags
    • last changes to service to prepare for shadow testing with production traffic; week 1 will be knocking out "necessary for testing", week two will be knocking out "necessary for production readiness" from this list: Feature Flags Rewrite: List of things to get done #22131 (comment)
  • helping with surveys support queue
  • Onboarding @haacked
  • Hackathon
• @havenbarnes
  • Results monitoring dashboard
  • embedding some existing early access feature functionality like opt-in user management, updating feature rollout phase manually etc (we should pause here for some feedback / early thoughts from customers imo)
  • configurable automated feature rollout phases
  • editable results monitoring dashboard
  • Templated rollout configs + templated results dashboards
  • Hackathon
• @haacked
  • First week - onboarding in-person in Seattle
  • Working on UX and SDK improvements (see onboarding ticket comment: https://github.com/PostHog/company-internal/issues/1650#issuecomment-2573711678)
  • Hackathon

Support

• 1/13-1/20 @dmarticus
• 1/20-1/27 @havenbarnes